Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with
or
.
Download ZIP
Commits on May 3, 2015
  1. @misto

    Merge pull request #82 from mlangc/public-val-renamed-only-locally-10…

    misto authored
    …02435
    
    Don't set `hasLocalScope` for public vals
Commits on May 2, 2015
  1. @mlangc

    Don't set `hasLocalScope` for public vals

    mlangc authored
    This commit fixes a bug in the implementation of
    `Rename.prepare(s: Selection).isLocalRename(t: global.Tree)` that caused
    `hasLocalScope` to be set for vals that might be accessed from other files.
    
    Extend the test suite, so that we can explicitly check for `hasLocalScope`
    
    Fixes #1002435 and #1002434
Commits on Mar 30, 2015
  1. @sschaef

    Merge pull request #80 from mlangc/class-renaming-1001928

    sschaef authored
    Take care of file rename operations in `rename` refactorings
Commits on Mar 29, 2015
  1. @mlangc

    Take care of file rename operations in 'rename' refactorings

    mlangc authored
    Prior to this patch it was up to clients of the refactoring library to figure
    out if source files should be renamed after a 'rename' refactoring. Doing so
    reliably however requires comparing the original ASTs to the refactored ones,
    which is not easily possible for clients. The approach that has been used so
    far by Scala IDE would only consider the tree where the refactoring has been
    initiated, which is broken by design as #1001928 illustrates. Thus rename
    operations are now supported directly by the refactoring library through
    `scala.tools.refactoring.common.RenameSourceFileChange`. Clients like Scala IDE
    can easily consummate his class and act accordingly.
    
    Adapt the unit tests to make file rename operations testable
    
    Code cleanups
    
    Fixes #1001928 in connection with an associated update for the IDE
Commits on Mar 16, 2015
  1. @misto

    Merge pull request #79 from mlangc/organize-imports-and-lazy-vals-100…

    misto authored
    …2402
    
    Properly handle type annotations on lazy vals when organizing imports
Commits on Mar 15, 2015
  1. @mlangc

    Handle type annotations on lazy vals properly

    mlangc authored
    Unfortunately the Scala compiler does not properly set range positions for lazy
    vals. We now work around this.
    
    Fixes #1002402
  2. @mlangc
  3. @misto

    Merge pull request #78 from mlangc/fix-renaming-lazy-vals-1002392

    misto authored
    Fix renaming for lazy vals
  4. @mlangc

    Prefer copying to modifying trees in place

    mlangc authored
    Modifying trees that are maintained by the Scala compiler does not seem to be
    such a good idea after all, even if we are only "fixing" them. Thanks to
    Simon Schäfer and Mirko Stocker for pointing this out.
  5. @mlangc

    Fix a cosmetic issue

    mlangc authored
  6. @mlangc

    Use clearer val names

    mlangc authored
  7. @mlangc

    Remove unnecessary cast

    mlangc authored
  8. @mlangc

    Fix renaming for lazy vals

    mlangc authored
    Unfortunately the compiler does not set modifier positions on
    DefDefs related to lazy vals. Therefore we now do this manually.
    
    Fixes #1002392
Commits on Mar 13, 2015
  1. @misto
Commits on Feb 23, 2015
  1. @misto

    Merge pull request #77 from mlangc/organize-imports-and-annotations-1…

    misto authored
    …001726
    
    Improves annotation handling for "Organize Imports"
Commits on Feb 22, 2015
  1. @mlangc

    Minor cleanup and refactoring

    mlangc authored
  2. @misto

    Merge pull request #76 from mlangc/support-multiple-scala-versions

    misto authored
    Support multiple scala versions
  3. @mlangc

    Documentation

    mlangc authored
  4. @mlangc

    Fix broken test case

    mlangc authored
    "scala.util.parsing.combinator.JavaTokenParsers" is not part of the standard library
    for Scala-2.11.
  5. @mlangc

    Improved annotations handling in "Organize Imports"

    mlangc authored
    The new implementation queries "AnnotationApi.tree" for additional information.
    As this method is not available for Scala-2.10, the improvements only apply to
    Scala-2.11.
    
    Fixes #1001726
  6. @mlangc

    Eclipse configuration update

    mlangc authored
    Add the source directory for Scala-2.11 to the Eclipse project.
Commits on Feb 21, 2015
  1. @mlangc

    Added support for multiple Scala versions

    mlangc authored
    The build conditionally includes "src/main/scala-2_10" or
    "src/main/scala-2_11", depending on the selected profile.
Commits on Feb 16, 2015
  1. @misto

    Merge pull request #75 from soc/patch-1

    misto authored
    Explicit creation of TermName
  2. @soc

    Explicit creation of TermName

    soc authored
    Create the TermName instance explicitly, because the implicit is deprecated.
Commits on Jan 18, 2015
  1. @misto

    Merge pull request #74 from mlangc/various-simple-bugfixes

    misto authored
    Treat locally defined types properly
  2. @mlangc
Commits on Jan 10, 2015
  1. @mlangc

    Treat locally defined types properly

    mlangc authored
    Fixes #73
Commits on Nov 4, 2014
  1. @misto

    Merge pull request #71 from retronym/ticket/69

    misto authored
    Reset reporter and typer before each unit test
  2. @retronym

    Reset reporter and typer before each unit test

    retronym authored
    Reusing the presentation compiler is good for performance
    but invites cross talk between tests. I modified
    `resetPresentationCompiler` to check if the reporter was
    clean before each test, and it turns out it wasn't. The
    typechecker predicates addition of some synthetic trees on
    an error-free reporter. I'm hoping this explains the errors
    we have seen in `add_override_final_flags_to_lazy_val` recently.
    
    Some code paths in the presentation compiler which run through
    `backgroundCompile` do reset the reporter. Perhaps we only see
    the problem sporadically (under load on Jenkins?) because we
    have to trigger a race condition that avoids the call to
    `backgroundCompile` before we pick up a work item in `pollForWork`.
    
    This commit explicitly resets the reporter before each test.
    It also does the same for the typer, although this is not in
    response to an observed bug, but rather just as a precaution.
    
    Relates to #69
Commits on Nov 3, 2014
  1. @misto

    Merge pull request #70 from sschaef/fix-lazy-check

    misto authored
    Don't call symbol.isLazy in test suite
  2. @sschaef

    Don't call symbol.isLazy in test suite

    sschaef authored
    This tests sometimes fails spuriously and it may be that the failure is
    triggered by the call of `isLazy`. While there is no guarantee that this
    change actually works around the problem, we have to try it.
    
    Fixes #69
Commits on Oct 29, 2014
  1. @misto
  2. @misto
  3. @misto
  4. @misto
Something went wrong with that request. Please try again.